Good morning, AI enthusiasts. Meta’s recruiting blitz just scored a major haul, with four OpenAI researchers jumping ship to Zuck’s new superintelligence team.

OpenAI CEO Sam Altman has shown confidence in retaining staff despite $100M offers, but Meta’s deep pockets are clearly talking — and its new unit is starting to take shape in a big way.

P.S. — Our next live workshop is today, June 27th, at 4 PM EST. Join and learn how to build advanced n8n agents that react, research, and automate like pros. RSVP here.

In today’s AI rundown:

• Meta poaches four OpenAI researchers

• Google’s Gemma 3n brings powerful AI to devices

• Convert lecture videos into detailed study materials

• Anthropic studies Claude’s emotional support

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

OPENAI & META

🥊 Meta poaches four OpenAI researchers

Image source: Midjourney / The Rundown

The Rundown: Meta has reportedly successfully recruited four OpenAI researchers for its new superintelligence unit, including three from OAI’s Zurich office and one key contributor to the AI leader’s o1 reasoning model.

The details:

• Zuckerberg personally recruited Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai, the trio that established OpenAI’s Zurich operations last year.

• Meta also landed Trapit Bansal, a foundational contributor to OpenAI's o1 reasoning model who worked alongside co-founder Ilya Sutskever.

• Sam Altman said last week that Meta had offered $100M bonuses in poaching attempts, but “none of OpenAI’s best people” had taken the offer.

• Beyer confirmed on X that the Zurich trio was joining Meta, but denied the reports of $100M signing bonuses, calling them “fake news”.

• Meta’s hiring spree comes after its $15B investment in Scale AI and poaching of its CEO Alexandr Wang to lead the new division.

Why it matters: Meta’s new superintelligence team is taking shape — and despite Altman’s commentary last week, at least four of his researchers are willing to make the move. With an influx of new talent from top labs and a clear willingness to spend at all costs, Meta’s first release from the new unit will be a fascinating one to watch.

TOGETHER WITH AIRIA

💼 The AI Platform for Enterprises

The Rundown: Airia is an enterprise platform that ensures secure creation, orchestration, and deployment of AI agents — enabling businesses to automate workflows and integrate systems with confidence, while maintaining robust security and compliance standards.

With Airia, you’ll experience:

• Built-in AI security safeguards for runtime prompts, data leakage, and prompt injections

• Scalable compliance that maintains regulatory standards (SOC 2, HIPAA, GDPR)

• Granular control and visibility for security policies, managing access control, and more

Book a Demo today.

GOOGLE

🚀 Google’s Gemma 3n brings powerful AI to devices

Image source: Google DeepMind

The Rundown: Google launched the full version of Gemma 3n, its new family of open AI models (2B and 4B options) designed to bring powerful multimodal capabilities to mobile and consumer edge devices.

The details:

• The new models natively understand images, audio, video, and text, while being efficient enough to run on hardware with as little as 2GB of RAM.

• Built-in vision capabilities analyze video at 60 fps on Pixel phones, enabling real-time object recognition and scene understanding.

• Gemma’s audio features translate across 35 languages and convert speech to text for accessibility applications and voice assistants.

• Gemma’s larger E4B version becomes the first model under 10B parameters to surpass a 1300 score on the competitive LMArena benchmark.

Why it matters: The full Gemma release is another extremely impressive launch from Google, with models continuing to get more powerful despite shrinking in size for consumer hardware. The small, open model opens up limitless intelligent on-device use cases.

AI TRAINING

🎓 Convert lecture videos into detailed study materials

The Rundown: In this tutorial, you will learn how to use the new video input feature of Google's Gemini to transform lecture videos into detailed notes and interactive quiz sessions to improve your study experience.

Step-by-step:

1. Go to Google's Gemini app and upload your lecture video.

2. Use this prompt: “Analyze this lecture video and provide: detailed outline, comprehensive notes, formulas/examples, and timestamps for each topic”

3. Follow up by requesting it to create a comprehensive quiz, plus answer keys with explanations

4. Ask it to code an interactive quiz based on this lecture content, and to include a hint button for when help is needed

Pro tip: Save all materials in one document and repeat this process for multiple lectures to build your complete course study library.

PRESENTED BY IBM

🚀 Lockheed Martin partners with IBM for AI

The Rundown: Lockheed Martin’s AI takeoff began with data cleanup — cutting its tools by 50% and replacing them with a single unified system that runs on IBM’s watsonx.data to drive faster, smarter engineering.

Learn how Lockheed Martin leveraged watsonx to:

• Help 10,00 engineers build scalable AI products in its AI Factory

• Respond to employee questions using AI, boosting accuracy by 20%

• Overcome data silos to create a single accessible, connected data environment

Read the full story.

ANTHROPIC

🫂 Anthropic studies Claude’s emotional support

Image source: Anthropic

The Rundown: Anthropic published new research on how Claude is used for emotional support and affective conversations, finding its use is far less common than reported, with companionship and roleplay accounting for under 0.5% of interactions.

The details:

• Researchers analyzed 4.5M Claude conversations using Clio, a tool that aggregates usage patterns while anonymizing individual chats.

• The data found that only 2.9% involved emotional support, with most focused on practical concerns like career transitions and relationship advice.

• Despite media narratives, the study showed that conversations seeking companionship or engaging in roleplay made up less than 0.5% of total use.

• Researchers also noted that users' expressed sentiment often grew more positive over the course of a chat, suggesting AI didn’t amplify negative spirals.

Why it matters: Recent media revealed some extreme cases of AI romance and dependency, but the data shows those are still few and far between (at least via Claude). However, Anthropic is dev-focused and less mainstream than ChatGPT or platforms like Character AI — so the numbers likely look a lot different elsewhere in AI.

QUICK HITS

🛠️ Trending AI Tools

• ⚙️ Gemini CLI - Open-source terminal agent with high free usage limits

• 📸 Higgsfield Soul - New high-aesthetic photo model with advanced realism

• 🧬 AlphaGenome - DeepMind’s new AI model for DNA analysis

• 🗣️ Voice Design V3 - Create any voice you can imagine with a prompt

💼 AI Job Opportunities

• 🧠 Mistral AI - Software Engineer

• 🚗 Waymo - Software Engineer

• 🏗️ Meta - Business Engineering Manager

• 🤖 Figure AI - Robotic Operations Technician

📰 Everything else in AI today

Black Forest Labs released FLUX.1 Kontext [dev], an open-weight, SOTA image editing model that can efficiently run on consumer hardware.

DeepSeek’s R2 model has faced issues due to export controls creating Nvidia chip shortages, with CEO Liang Wenfeng not happy with the model’s performance.

OpenAI released a series of updates, including Deep Research via API, Web Search in o3 and o4-mini, and its next DevDay event, slated for Oct. 6 in San Francisco.

HeyGen introduced HeyGen Agent, a “Creative Operating System” that creates video content with scripts, actors, edits, and more from a simple text, image, or video.

Google launched Doppl, a new experiment on its Labs platform, allowing users to create AI-generated try-on videos from a photo and a product.

Meta became the latest AI company to earn a favorable “fair use” ruling in court, winning a lawsuit brought by authors over copyright infringement.

Suno announced the acquisition of WavTool, bringing the startup’s browser-based digital audio workstation to the platform for more advanced music creation.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop today, June 27th at 4 PM EST with Dr. Alvaro Cintas, The Rundown’s AI professor. By the end of the workshop, you’ll confidently be able to build advanced n8n agents that react, research, and automate like pros.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Good morning, robotics enthusiasts. Google DeepMind has unveiled its first vision-language-action model that runs entirely onboard a robot — no cloud, no cables, just self-contained intelligence.

It’s part of a new wave of untethered robotics — machines that can see, think, and act without phoning home. The promise? Smarter, faster, and more autonomous robots, ready to take on the real world.

In today’s robotics rundown:

• Google DeepMind’s on-device robot AI

• Apptronik’s ‘superhuman’ robotics spin-off

• MIT fast-tracks robot design with generative AI

• Eco-friendly robots made from rice paper

• Quick hits on other robotics news

LATEST DEVELOPMENTS

GOOGLE DEEPMIND

🧠 Google DeepMind’s on-device robot AI

Image source: Google DeepMind

The Rundown: Google DeepMind just unveiled Gemini Robotics On-Device, its first vision-language-action (VLA) model built to run entirely on a robot — with no internet needed. And it delivers nearly the same muscle as its cloud-powered counterpart.

The details:

• The new model allows robots to interpret what they see, understand instructions, and carry out complex physical tasks entirely on their own. 

• Google says that despite its compact size, the model’s performance is nearly on par with the cloud-based Gemini Robotics system.

• It can adapt to new tasks with as few as 50 to 100 demos, making an ideal solution for rapid deployment in changing environments.

• Google trained the model on its ALOHA robot, but says it is adaptable to different robot types, including humanoids and bi-arm robots.

Why it matters: As robot models like 1X’s Redwood and Figure AI’s Helix expand into varied physical environments, Google is charting a different course — prioritizing flexibility. Its Gemini-based robotics system, designed to run both in the cloud and locally, will enable teams to build latency-sensitive apps, even in fully offline settings.

APPTRONIK

🤖 Apptronik’s ‘superhuman’ robotics spin-off

Image source: Apptronik

The Rundown: Austin-based Apptronik, makers of the Apollo humanoid, launched a new subsidiary called Elevate Robotics — but instead of industrial humanoids, their focus is on “superhuman” machines that aren’t limited by the human form.

The details:

• The new subsidiary will build machines designed for heavy-duty industrial tasks, such as lifting massive loads and operating in extreme conditions.

• This strategic spin-off comes hot on the heels of Apptronik’s $403M funding round, led by B Capital and Capital Factory, with participation from Google.

• Driven by Plus One Robotics co-founder Paul Hvass, Elevate inherits Apptronik’s decade of breakthroughs, from exoskeletons to robotic arms.

• That said, unlike Apollo, Elevate’s robots will not be limited to the two-legged, human-like design, allowing for more specialized and robust architectures.

Why it matters: Elevate promises multipurpose automation robotics designed for extreme loads, relentless precision, and environments too dangerous or demanding for people — or even humanoids — to handle. The move signals a new wave of automation that might leave bipedal industrial humanoids in the dust.

MIT

🔥 MIT fast-tracks robot design with generative AI

Image source: MIT’s CSAIL

The Rundown: MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) is leveraging generative AI to fundamentally change how robots are conceived and built — turning a historically slow, manual process into a fast and automated one.

The details:

• In two new projects, MIT used an AI platform to test thousands of underwater robot designs — including jumpers and gliders — inside a physics simulator. 

• The AI optimized every parameter, from fin shape to weight, and spat out the best designs — a task that would have taken months of trial-and-error.

• Once the AI crowned its champions, CSAIL’s team 3D-printed the winning blueprints and threw them into real-world testing.

• The jump-bot shattered depth records, propelling itself off ocean floors with biomimetic precision, while the glider sliced through water with minimal drag.

Why it matters: Critically, these weren’t digital fantasies: the physical prototypes performed within 5% of their simulated predictions, validating the AI’s ruthless efficiency. This workflow — AI-driven design, simulation torture-testing, and then instant fabrication — could make robotics’ slow-motion R&D loops a thing of the past.

UNIVERSITY OF BRISTOL

🍚 Eco-friendly robots made from rice paper

Image source: University of Bristol

The Rundown: Scientists at the University of Bristol just found that Vietnamese spring roll wrappers work as well as silicone materials in prototyping flexible robots — plus, they’re nontoxic and biodegrade completely within a month.

The details:

• Unlike traditional plastics or silicones, this kitchen material is fully biodegradable, allowing soft robots to be composted after use.

• Because it’s edible, the material is completely nontoxic, making it ideal for applications in medical, educational, or food-handling environments.

• The team was able to shape it into grippers, actuators, and even sensors that mimic the softness and adaptability of biological tissues.

• Its natural transparency also allows for easy integration with optical sensors, while its malleability lets researchers customize thickness and texture.

Why it matters: Bristol’s rice paper robot turns a humble kitchen staple into a biodegradable, non-toxic wonder, making soft robotics accessible, affordable, and eco-friendly. Like similar robotics projects that use starch or algae, it’s a recipe for rapid prototyping, safer outreach, and single-use applications that won’t leave a trace.

QUICK HITS

📰 Everything else in robotics today

Waymo and Uber launched robotaxi rides in Atlanta, marking Waymo’s fifth major U.S. market and the second city under its deal with Uber after Austin.

Tesla CEO Elon Musk teased new progress on the company’s upcoming humanoid, Optimus V3, and confirmed that it is already using a Grok voice assistant.

Software firm Nominal raised $75M to expand its unified platform that accelerates hardware development and testing for aerospace, defense, and robotics.

Melco Mobility, a Mitsubishi spin-off, ordered nearly 100 industrial robots from U.S.-based Cartken to automate material handling operations throughout Japan.

Ontario introduced Canada’s first beach-cleaning robot, a fully electric rover known as BeBot, to remove litter from provincial park shorelines.

Cobot leader Universal Robots unveiled UR Studio, a next-gen online simulation tool designed to transform how robotic work cells are customized and deployed.

Bike-sharing firm Hellobike moved into the robotaxi sector by forming a new joint venture with tech powerhouse Ant Group and global battery giant CATL.

ROKAE Robotics, a robotics firm based in Beijing, has unveiled two new industrial humanoid models for factory jobs.

Galbot, a Chinese startup developing general-purpose AI robots, raised $152M, bringing total funding to $332M in just two years.

California-based K-Scale Labs announced that it will manufacture its $9K open-source humanoid robots in Dallas.

German startup NEURA Robotics unveiled its 4NE1 Gen3 at Automatica 2025 in Munich as it is reportedly seeking to raise up to $1.2B in new funding.

Northern Arizona University researchers unveiled OpenExo, the world’s first open-source robotic exoskeleton framework, to accelerate progress in the field.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Friday, June 27th, at 4 PM EST with Dr. Alvaro Cintas, The Rundown’s AI professor. By the end of the workshop, you’ll be able to build advanced n8n agents that react, research, and automate like pros.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Jennifer, and Joey—The Rundown’s editorial team

Good morning, AI enthusiasts. Google DeepMind just compressed decades of genetic research into a four-hour training run, and the result could change our understanding of disease.

With the new AlphaGenome model quickly predicting mutation effects across 1M-letter DNA sequences on GPUs instead of in the lab, the ability to pinpoint diseases at their genetic source is suddenly within reach.

In today’s AI rundown:

• DeepMind’s AlphaGenome for DNA analysis

• Google drops open-source Gemini CLI

• Build your AI assistant with automated workflows

• Anthropic adds app-building capabilities to Claude

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

GOOGLE DEEPMIND

🧬 DeepMind’s AlphaGenome for DNA analysis

Image source: Google DeepMind

The Rundown: Google DeepMind just released AlphaGenome, a new AI model that predicts how DNA mutations affect thousands of molecular processes by analyzing sequences up to 1M base-pairs long.

The details:

• The model reads DNA stretches 100x longer than older tools, predicting how nearby genes will behave and how other regulatory regions function.

• The release unifies thousands of molecular predictions into one tool, while still beating out most specialized models across a range of genomic benchmarks.

• Researchers tested it on leukemia patients, helping identify how specific mutations switched on cancer-causing genes that should have stayed silent.

• DeepMind trained the entire system in just four hours using public genetic databases, consuming half the computing power of their previous DNA model.

Why it matters: AlphaGenome moves complex biological research from the lab to the computer, letting scientists test genetic hypotheses at an unprecedented scale. While not a crystal ball for personal health, it gives researchers a powerful first guess, dramatically speeding up the search for mutations and variants that cause disease.

TOGETHER WITH SANA

🤖 Meet your AI operating system

The Rundown: Sana Agents is where teams go to work with AI. Co-create flawless reports, decks, and dashboards — automatically built from your live company data.

With Sana Agents, you can:

• Work inside one hub that merges search, notes, canvases, workflows, and agents.

• Hand off docs, slides, sheets, and tickets to AI that drafts and updates them end-to-end

• Get precise answers grounded in your data and wrapped in enterprise-grade security

Drive enterprise-wide innovation today with Sana Agents.

GOOGLE

⚙️ Google drops open-source Gemini CLI

Image source: Google

The Rundown: Google just released Gemini CLI, a new open-source terminal agent that brings Gemini 2.5 Pro directly to developers' command lines with high free usage limits.

The details:

• Developers get 60 requests per minute and 1,000 daily queries at no charge, limits that Google set after doubling its own internal usage patterns.

• The Apache 2.0 licensed tool supports Model Context Protocol, bundled extensions, and custom GEMINI.md files for project-specific configurations.

• Other built-in capabilities include Google Search grounding, file manipulation, command execution, and Imagen/Veo integration for multimedia generations.

• CLI is integrated directly with Code Assist, leveraging Gemini 2.5 Pro and its 1M context window — currently the highest ranked model on the WebDev Arena.

Why it matters: The battle for AI developer adoption just got even more competitive, with Google weaponizing free access against OpenAI and Anthropic’s paid competitors. With generous limits and the open-source move addressing enterprise security concerns, Google hopes to shift dev workflows entirely into its ecosystem.

AI TRAINING

🤖 Build your AI assistant with automated workflows

The Rundown: In this tutorial, you will learn how to create a personal AI assistant that can draft emails, manage tasks, and handle requests using n8n's ability to connect AI models with your favorite apps.

Step-by-step:

1. Head over to n8n and create a new workflow with a chat trigger

2. Add an AI Agent with your preferred model (GPT-4, Claude) and a Simple Memory

3. Connect your apps (Gmail, Calendar, Slack) and set parameters to “define automatically by the model”

4. Add custom instructions: “You are a helpful assistant with access to email and calendar tools. Use them when users make requests.”

Pro tip: Start with email drafting and calendar management, then expand to document creation and data analysis as you get comfortable.

PRESENTED BY IBM

📊 Data solutions to AI problems

The Rundown: IBM understands the challenges your organization faces in getting data ready for AI, and the company’s new “Data Matters” chapter breaks down how to turn scattered, messy information into high-octane fuel for models.

Inside, you’ll discover how to:

• Turn structured & unstructured data from a variety of sources into high-quality data for AI

• Overcome barriers to data preparedness

• Create a consistent methodology to assess the quality of your data

Click here to download the E-Book.

ANTHROPIC

🚀 Anthropic adds app-building capabilities to Claude

Image source: Anthropic

The Rundown: Anthropic upgraded Claude with new app-building capabilities, allowing any user to create, host, and share interactive AI-powered apps directly from simple text prompts via its “Artifacts” workspaces.

The details:

• Users can now build functional tools like data analyzers or study aids simply by describing their idea, with Claude handling the writing of all underlying code.

• The platform shifts costs to end users who sign in with Claude credentials, eliminating the need for creators to manage API keys or pay for others' usage.

• Free tier users can create and share apps while Pro ($20/month) and Team ($25-30/month) subscribers unlock advanced features and higher usage limits.

• Anthropic said over 500M artifacts have been created since launch last August, now with a dedicated space on the platform for better organization.

Why it matters: While Claude’s models are still highly preferred by developers for coding tasks, Anthropic’s Artifacts also help continue the vibe coding push — with smooth UI and ecosystem upgrades that push users to keep building naturally regardless of developer experience.

QUICK HITS

🛠️ Trending AI Tools

• 🗣️ ElevenLabs App - Powerful AI voice tools, now available on mobile

• ✈️ Airtable AI - Build enterprise apps that automate and manage operations

• 🤖 Gemini Robotics - VLA model that works locally on robots, without internet

• 🛡️ XBOW - Boost offensive security with autonomous AI

💼 AI Job Opportunities

• 🎨 The Rundown - Designer (Brand & Platform)

• 🛡️ Horizon3 - Senior Product Manager

• 🗣️ Rad AI - Senior ML Research Scientist

• 💼 Perplexity AI - Finance Manager

📰 Everything else in AI today

Postman launched a new AI-Readiness Hub with a 90-day plan and dev toolkit to help make your APIs agent-ready.*

Higgsfield AI released Soul, a new “high-aesthetic” photo model with advanced realism and 50+ presets for easy style optimization.

Creative Commons unveiled CC Signals, a new opt-in metadata system for dataset owners to spell out exactly how AI models may reuse their work.

ElevenLabs introduced Voice Design v3, featuring new upgrades for more expressive voice outputs and support for over 70 languages with accurate accents.

OpenAI released new Connectors for Pro ChatGPT accounts, giving users the ability to integrate data from tools including Google Drive, Dropbox, SharePoint, and Box.

Getty dropped its lawsuit against Stability AI that accused the company of copyright theft, following a “fair use” ruling in a separate case by authors against Anthropic.

Amazon announced new AI features for its Ring home security systems, including AI-generated video descriptions that provide users with real-time text updates.

*Sponsored Listing

COMMUNITY

🎥 Join our next live workshop

Join our next workshop this Friday, June 27th, at 4 PM EST with Dr. Alvaro Cintas, The Rundown’s AI professor. By the end of the workshop, you’ll be able to confidently build and run your own automation agents using n8n.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Good morning, AI enthusiasts. AI’s fair use debate just got its first major answer — and it’s both a win and a warning for AI companies.

With a judge green-lighting Anthropic’s use of legally purchased data but slamming them for using over 7M pirated books, the split decision might be a strong (but costly) precedent for AI’s legal woes.

Reminder: Our next live workshop is today at 3 PM EST — join and learn how to build production-grade AI agents that combine context, reasoning, and actions with Gumloop. RSVP here.

In today’s AI rundown:

• Anthropic scores win over AI ‘fair use’ claim

• OpenAI’s Workspace, Office competitor

• How to automate content strategy with scheduled tasks

• LinkedIn co-founder bets on AI ultrasound helmet

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

ANTHROPIC

📚 Anthropic scores win over AI ‘fair use’ claim

Image source: Reve / The Rundown

The Rundown: A federal judge just ruled that Anthropic's AI training on legally purchased books qualifies as fair use, but rejected any defense for 7M pirated copies in its digital library in the company’s class action lawsuit filed by a group of authors.

The details:

• The judge called AI training "spectacularly" transformative, comparing Claude to aspiring writers learning from established authors rather than copying them.

• The authors failed to demonstrate that Claude could generate outputs resembling their original works, weakening core claims about competitive harm.

• The filings revealed that Anthropic legally spent “many millions” to purchase print books, scanning them into digital files for use in AI training.

• However, Anthropic also downloaded millions of books from pirate sites, storing them permanently, which the court said violated authors’ rights.

• The company will face trial in December for willful infringement of the pirated works, with potential damages potentially reaching $150,000 per book.

Why it matters: AI labs get a green light to train on legally obtained data, which could provide one of the first precedents for countless cases currently levied against tech companies. However, given the lack of clarity around how many copyrighted materials get into training data, this is still just one battle in what looks like a very long legal war.

TOGETHER WITH H COMPANY

🤖 The AI agent that completes tasks for you

The Rundown: Backed by a historic $220M seed round, H Company just released their flagship Runner H AI agent that lets you automate entire tasks in a few seconds — now available in beta.

Some examples of tasks you can delegate to Runner H:

• Reading your important emails and drafting (or even sending!) replies

• Creating a Google Sheets with trending ad ideas and sharing it with your team on Slack

• Finding job opportunities and applying for you

Offload some work to Runner H today — try it now for free.

OPENAI

📊 OpenAI’s Workspace, Office competitor

Image source: o3 / The Rundown

The Rundown: OpenAI is reportedly building productivity tools for ChatGPT that mirror Google Workspace and Microsoft Office — with features like real-time document collaboration and multi-user chat.

The details:

• OpenAI CPO Kevin Weil reportedly first showcased collaboration designs last year, though development stalled until the Canvas interface launch in October.

• The Information reports that OpenAI has built but has yet to release multiuser chat, allowing teams to communicate about shared work directly in ChatGPT.

• OpenAI also recently rolled out a record mode for transcriptions, file uploads to Projects, and connectors to pull data from Teams, Drive, and DropBox.

• Business subscriptions generated $600M in 2024, with OpenAI projecting $15B by 2030, with increased revenue coming from enterprise subscriptions.

Why it matters: Sam Altman warned last year that OpenAI would “steamroll” most AI startups… But he might also have his biggest partner in the crosshairs. ChatGPT’s productivity push is about to step right on Microsoft’s legacy software — and given the icy current relationship, the renegotiation may get even more contentious.

AI TRAINING

🤖 How to automate your content strategy with scheduled tasks

The Rundown: In this tutorial, you’ll learn how to use Grok’s new Tasks feature to schedule automated content research and trend analysis — running silently in the background and notifying you once it's done.

Step-by-step:

1. Visit Grok and navigate to the “Tasks” section

2. Click “Add new” and create your first task: “Weekly Content Trends Analysis” scheduled for Sundays

3. Use this prompt: “Analyze trending topics in [your niche] from the past week. Identify 5 content opportunities with content format recommendations”

4. Create additional tasks such as viral content analysis, or monthly market research (up to 10 total tasks)

Pro tip: Enable DeepSearch for better results. When tasks are complete, review notifications and refine prompts based on results quality.

PRESENTED BY IBM

🎙️Transforming HR with agentic AI

The Rundown: Curious about agentic AI in HR? Steve Moss, Director of watsonx Americas at IBM, discusses HR, agentic AI, and the real-world impact of IBM’s transformation journey in the latest episode of the AI in Action podcast.

Dive into the details of how IBM has:

• Used agentic AI and consolidated various HR tools into a single interface

• Improved employee experience and operational efficiency

• Driven strategic shifts, boosted productivity, and scaled teams effectively

Click here to watch the video.

AI & NEUROTECHNOLOGY

🧠 LinkedIn co-founder bets on AI ultrasound helmet

Image source: o3 / The Rundown

The Rundown: LinkedIn co-founder and OpenAI investor Reid Hoffman just led a $12M funding round for Sanmai Technologies, which is developing AI-guided ultrasound devices for treating mental health conditions without surgery.

The details:

• Sanmai’s consumer devices focus ultrasound waves on specific brain regions to treat anxiety, depression, and enhance cognitive function.

• The startup combines the ultrasound tech with AI coaching systems into a helmet at a sub-$500 price point, targeting consumers’ in-home use.

• Hoffman joined Sanmai's board through his Aphorism Foundation, saying non-invasive approaches are “much less risky” than tech like Neuralink.

• The company is currently testing anxiety treatments with a prototype at its Sunnyvale facility ahead of FDA trials.

Why it matters: Tech billionaires like Elon Musk, Jeff Bezos, Bill Gates, and now Reid Hoffman are all funding brain-tech startups. With an AI coach guiding treatments and a non-invasive approach, the start of the neurotech wave may end up being a lighter touch that is easier to swallow for consumers than a full brain-computer interface.

QUICK HITS

🛠️ Trending AI Tools

• 💰 Taka - Your AI Personal CFO: Ask any personal finance question. Make confident decisions with a self-building canvas tailored to you*

• 🗣️ 11ai - ElevenLabs’ new voice assistant with MCP integrations

• 💻 Mu - Microsoft’s fast, local AI for Windows Copilot + PCs

• 🎆 Imagen 4 Ultra - Google’s top image model, available via API & AI Studio

  *Sponsored Listing

💼 AI Job Opportunities

• 🔁 The Rundown - Strategic Partnerships (AI University)

• 🤝 OpenAI - Program Manager, M&A Recruiting

• 🖥️ Together AI - Senior Systems Administrator

• 📦 Groq - Senior Program Manager

📰 Everything else in AI today

SimilarWeb data shows ChatGPT downloads on iOS hit 29M+ over the last 28 days, nearly surpassing downloads of TikTok, Facebook, and Instagram (33M) combined.

Sam Altman said that the ‘io’ lawsuit is “silly, disappointing and wrong", saying that founder Jason Rugolo made persistent attempts to get acquired by OpenAI.

Mira Murati’s Thinking Machines Lab is planning to develop custom AI models to help businesses increase profits, according to a new report from The Information.

Google released Gemini Robotics On-Device, a new VLA model that powers robotics dexterity and task completion without needing an internet connection.

Databricks & Perplexity co-founder Andy Konwinski launched the Laude Institute, pledging $100M to fast-track computer science breakthroughs for real-world impact.

XBOW revealed that its autonomous AI became the first to surpass all humans on the HackerOne platform, also announcing a new $75M Series B funding round.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop today at 3 PM EST with Max Brodeur-Urbas, Founder & CEO at Gumloop. By the end of the workshop, you’ll know how to build production-grade AI agents that combine context, reasoning, and actions with Gumloop.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason — The Rundown’s editorial team

Good morning, tech enthusiasts. The Vera Rubin Observatory has just released its first cosmic images—captured by a digital camera so powerful it can spot a golf ball from 15 miles away.

Over the next decade, this observatory will scan the night sky repeatedly, uncovering hidden asteroids and distant galaxies. The question is: What new cosmic secrets might unfold when we view the universe in Ultra-HD?

In today’s tech rundown:

• World’s largest camera reveals deep space

• Murati’s Thinking Machines Lab nabs $2B

• China on the verge of ‘100 Deepseeks’

• Tesla robotaxis hit regulatory speed bump

• Quick hits on other major tech news

LATEST DEVELOPMENTS

SPACE TECH

🔭 World’s largest camera reveals deep space

Image source: Vera C. Rubin Observatory

The Rundown: The Vera C. Rubin Observatory, perched on Chile’s Cerro Pachón, just unveiled the first images from its Legacy Survey of Space and Time (LSST) camera — the world’s largest digital camera, with a staggering 3,200-megapixel sensor.

The details:

• These initial shots, generated from a mosaic of 678 exposures during a brief test run, showed millions of galaxies, stars, and unknown asteroids.

• The LSST camera, roughly the size of a small car, is a marvel of modern engineering, designed to capture more detail than any previous telescope.

• Its ultra-wide lens can capture an area 45 times the size of the full moon in a single exposure.

• The LSST will run for 10 years, repeatedly photographing the entire southern sky every three to four nights, creating a time-lapse view of the cosmos.

Why it matters: Researchers now will be able to track everything from near-Earth asteroids to supernovae and the movements of dark matter — without the usual bottleneck of telescope time applications. As the observatory ramps up, the sheer volume and quality of its data promise to transform our understanding of the universe.

THINKING MACHINES LAB

💰 Murati’s Thinking Machines Lab nabs $2B

Image source: Atomico/YouTube

The Rundown: Thinking Machines Lab — the secretive startup founded by OpenAI’s former CTO Mira Murati — just raised a staggering $2B at a $10B valuation, marking what may be the largest seed-stage funding ever recorded.

The details:

• The round was led by heavyweight VCs Andreessen Horowitz, with participation from Conviction Partners and other top-tier firms.

• Murati, who led the development of ChatGPT and DALL-E, heads a team of about 30 engineers, nearly two-thirds of whom are former OpenAI employees.

• With no public product yet, Thinking Machines Lab is banking on the reputation and vision of its star team to attract capital and talent.

• The startup aims to solve “black box” AI issues, prioritizing transparency and user control, a departure from opaque models that dominate the field.

Why it matters: For now, the startup’s astronomical valuation is less about what it’s built than what it might build — a bet that this team, with its track record of delivering groundbreaking AI, can once again redefine the field. In a sector where talent is king, Thinking Machines Lab looks to be a major player in AI’s next evolution.

CHINA AI

🤯 China on the verge of ‘100 DeepSeeks’

Image source: DeepSeek

The Rundown: China is on the verge of a major AI boom, with a former top official predicting that the country will see more than 100 DeepSeek-like AI breakthroughs within the next 18 months, according to Bloomberg.

The details:

• Zhu Min, previously a deputy governor of the People’s Bank of China, made these remarks during the World Economic Forum in Tianjin.

• He emphasized that these new software innovations “will fundamentally change the nature and the tech nature of the whole Chinese economy.”

• The forecast is driven by China’s deep reservoir of engineering expertise, its enormous consumer market, and massive government investments in tech.

• DeepSeek has become a poster child for Chinese AI ingenuity, offering models that rival OpenAI’s best at a fraction of the price.

Why it matters: As Zhu’s prediction suggests, DeepSeek is just the beginning — if he’s right, we’re on the verge of a tidal wave of AI innovation from China. And this is all after semiconductor sanctions from the U.S., which are forcing them to operate at lower costs and under tight constraints.

TESLA

🚕 Tesla’s robotaxis hit regulatory speed bump

Image source: Tesla

The Rundown: Just days after Tesla launched its robotaxi service in Austin, the U.S. National Highway Traffic Safety Administration (NHTSA) is reportedly investigating them over viral videos showing the vehicles behaving erratically on public roads.

The details:

• Tesla just launched a limited rollout with 10 to 20 Model Ys equipped with the company’s latest FSD Unsupervised software and hardware.

• Social media clips captured the robotaxis making erratic maneuvers, including entering oncoming traffic and stopping abruptly in the middle of roads.

• NHTSA confirmed it is “aware of the referenced incidents and is in contact with the manufacturer to gather additional information.”

• The launch — which includes a human valet in the front passenger seat ready to intervene if needed — sparked an 8% spike in Tesla’s share price.

Why it matters: Tesla’s robotaxi launch arrives amid intense regulatory scrutiny, with the NHTSA already probing its Full Self-Driving tech after several serious crashes. For now, Waymo maintains a clear lead in the space, operating a large commercial fleet with over 250K weekly rides across multiple cities and a proven safety record.

QUICK HITS

📰 Everything else in tech today

New York Governor Kathy Hochul announced plans to build a zero-emission nuclear power plant in the state to meet the demands of AI and semiconductor production.

The UK government plans to allocate more than £500M to boost domestic quantum-computing research, according to the Financial Times.

Tesla said it has invested about $44B in U.S. manufacturing and infrastructure since its founding, with roughly $10B of that total deployed in the last fiscal year alone.

Andy Konwinski, co-founder of Databricks and Perplexity, is forming a new AI research institute backed by $100M of his personal funds through his company Laude.

SoftBank founder Masayoshi Son is reportedly in discussions to set up a $1T industrial complex in Arizona that will build robots and AI.

Midjourney launched its first AI video model just a week after Disney and NBCUniversal filed a copyright-infringement lawsuit against the startup.

Cluely, the startup that promises to help users “cheat on everything,” announced a $15M Series A funding round led by Andreessen Horowitz.

An MIT study shows that overusing AI tools such as ChatGPT can erode memory and reduce critical thinking abilities.

AI design software company Canva is organizing a $400–500M secondary share sale at a $37B valuation, according to The Information.

Honda's research subsidiary successfully tested a reusable rocket in Japan, achieving its first landing after going up to 890 ft.

Legal AI company Harvey raised $300M at a $5B valuation, counting top law firms like Paul, Weiss, and major corporate legal departments among its clients.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop on Wednesday, June 25th at 3 PM EST with Max Brodeur-Urbas, Founder & CEO at Gumloop. By the end of the workshop, you’ll know how to build production-grade AI agents that combine context, reasoning, and actions with Gumloop.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Jennifer, and Joey—The Rundown’s editorial team

Good morning, AI enthusiasts. Jony Ive and Sam Altman’s recent ‘io’ hardware partnership just got scrubbed from the internet — courtesy of a court order from a startup claiming the AI giant stole more than just a name.

With Google X spinoff iyO alleging suspicious meetings and an eerily similar concept, is the power duo’s hyped AI device more imitation than innovation?

In today’s AI rundown:

• OpenAI scrubs 'io' over trademark clash

• ElevenLabs debuts new voice assistant

• How to optimize prompts for better AI output

• Reddit eyes Altman’s World ID for human verification

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

OPENAI

⚖️ OpenAI scrubs 'io' over trademark clash

Image source: iyO

The Rundown: OpenAI just removed all promotional materials for its $6.5B acquisition of Jony Ive's AI hardware startup, io, reportedly due to a court order tied to a trademark dispute with Google X spinout iyO.

The details:

• iyO creates hardware that “allows users to interact with their smartphones, computers, AI, and the internet without the use of physical interfaces.”

• The startup’s latest product is the iyO One AI-powered earbuds, a “computer without a screen” that can run apps via voice and converse with the user.

• The filing alleges that Sam Altman and Ive’s LoveFrom met with iyO initially in 2022 and again in Spring 2025 just before the io announcement.

• OpenAI removed the blog post and nine-minute video featuring Ive and Altman from its website and YouTube channels following the legal action.

• However, the company maintains that the acquisition remains on track, calling the trademark complaint "utterly baseless."

Why it matters: While this is unlikely to derail OpenAI and Jony Ive’s hardware plans, the alleged details in the filing — from LoveFrom employees purchasing the device and directly asking to share details to the nearly identical name itself — certainly don’t paint the same picture of innovation the AI leader has created thus far around io.

TOGETHER WITH HUBSPOT

💰 Turn AI into your income engine

The Rundown: HubSpot’s new “200+ AI-Powered Income Ideas” free guide offers actionable strategies to turn artificial intelligence into your own personal revenue generator — unlocking a gateway to financial innovation in the digital age.

With this guide, you can:

• Explore hundreds of revenue-generating ideas across industries with real-world applications

• Follow simple, step-by-step instructions that make AI accessible to everyone

• Adopt cutting-edge strategies to keep you ahead in today's fast-paced market

Access your free guide here.

ELEVENLABS

🗣️ElevenLabs debuts new voice assistant

Image source: ElevenLabs

The Rundown: AI voice platform ElevenLabs just launched 11ai, an in-house voice assistant that connects to tools via Anthropic’s Model Context Protocol to execute tasks rather than just answer questions.

The details:

• The experimental alpha release integrates with platforms like Perplexity, Linear, Slack, and Notion, allowing users to manage tasks through voice commands.

• Developers can also add custom MCP servers for additional integrations and workflows beyond the pre-built connections.

• The platform offers 5,000+ voice options and supports voice cloning, running on ElevenLabs’ own conversational AI infrastructure.

• 11ai is free to use and try for “several weeks,” allowing the company to gather feedback on the product and integrations.

Why it matters: ElevenLabs already has some of the strongest speech models on the market, and pairing them with MCP for tools, data access, and actions could showcase how voice assistants can actually be more useful than Siri and other outdated assistants may have led consumers to believe.

AI TRAINING

🔧 How to optimize prompts for better AI output

The Rundown: In this tutorial, you will learn how to use OpenAI Playground’s new automatic prompt optimization tool to transform basic prompts into high-performance system messages for more effective AI interactions.

Step-by-step:

1. Go to OpenAI Playground and access the Prompts section

2. Write your basic system message describing what you want the AI to do

3. Click the “Optimize” button to automatically improve your prompt with better structure and clarity

4. Review it and then “Save” it with a descriptive name to reuse in projects and API calls

Pro tip: Test optimized prompts with various inputs to ensure consistent performance across different use cases.

PRESENTED BY BASETEN

⚙️ Fine-tune without limits

The Rundown: Baseten just launched Training in beta, providing the infrastructure to fine-tune and train without timeout limits using any model or dataset. Bring your scripts and data, Baseten will handle the rest.

What you get with Training:

• On-demand compute — pay only for what you use

• No job limits or vendor lock-in

• Out-of-the-box checkpoints and caching

Request access to Baseten Training.

REDDIT

👁️ Reddit eyes Altman’s World ID for human verification

Image source: o3 / The Rundown

The Rundown: Reddit is reportedly negotiating with Sam Altman’s Tools For Humanity to integrate the company’s iris-scanning World ID Orb system, allowing users to provide proof of humanity while staying anonymous.

The details:

• The system would offer Reddit users optional verification through World ID’s encrypted iris scans, which fragment biometric data across servers worldwide.

• CEO Steve Huffman hinted at the shift last month, posting on efforts to preserve anonymity while deterring the flood of AI accounts on the platform.

• World ID assigns users cryptographic proof without storing personal data, though minors under 18 are currently blocked from the Orb scanning process.

• The partnership would position Reddit as the first major U.S. social platform to test biometric verification at scale aside from simple email checks.

Why it matters: The “Dead Internet” theory of the web becoming overrun with AI bots is a real concern — and something already being experienced across social media platforms. While World’s Iris scanning initiatives were initially met with tons of skepticism and anger, the need for human verification is going to be very real.

QUICK HITS

🛠️ Trending AI Tools

• ⚡ Deepgram Voice Agent API - Build production-ready voice agents with a unified speech-to-speech API*

• 📖 Kimi-Researcher - Moonshot AI’s new SOTA research agent

• 🗣️ Voice Design - MiniMax’s customizable, multilingual voice generator

• 🤏 Mistral Small 3.2 - Updated with better instruction following, fewer errors

*Sponsored Listing

💼 AI Job Opportunities

• 📞 The Rundown - Account Manager

• 🌍 Cresta - Product Manager

• 🔬 Anthropic - Research Manager

• ⚖️ OpenAI - Product Counsel

📰 Everything else in AI today

Disney has been in talks with “companies like OpenAI” to license its characters and IP, but its lawsuit against Midjourney “likely won’t be the last” against AI firms.

A U.S. official claimed DeepSeek is working with the Chinese government on military and intelligence ops, while using workarounds to access advanced AI chips.

Google released Magenta RealTime, an open live music AI and “cousin” of its Lyria model, allowing users to create/blend music live and locally on consumer hardware.

Meta also met with Runway for a potential acquisition, in addition to reported meetings with SSI, Thinking Machines, and Perplexity, though a deal never materialized.

Softbank CEO Masayoshi Son is reportedly pitching a “Crystal Land” $1B megahub for AI and robotics manufacturing in Arizona, courting TSMC and Samsung.

Microsoft debuted Mu, a new small language model that powers agentic capabilities in Settings for on-device use on Windows Copilot + PCs.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop on Wednesday, June 25th at 3 PM EST with Max Brodeur-Urbas, Founder & CEO at Gumloop. By the end of the workshop, you’ll know how to build production-grade AI agents that combine context, reasoning, and actions with Gumloop.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team

Good morning, robotics enthusiasts. UK researchers have engineered an artificial skin that’s soft, flexible, and doesn’t rely on embedded sensors.

This skin can ‘feel’ pressure, heat, and even pain across its entire surface. The result? A leap toward robots and prosthetics that can respond to their environment more like we do.

In today’s robotics rundown:

• New artificial robot skin can ‘feel’ pain

• Softbank eyes $1T robotics complex in Arizona

• Germany’s Neura to raise $1.15B for humanoids

• China’s mosquito-sized drone for military

• Quick hits on other robotics news

LATEST DEVELOPMENTS

UNIVERSITY OF CAMBRIDGE

👉🏼 New artificial robot skin can ‘feel’ pain

Image source: University of Cambridge

The Rundown: University of Cambridge and University College London researchers just developed a stretchy artificial skin that allows robots to detect pressure, temperature, and maybe even pain.

The details:

• This next-gen skin is made from gelatin-based hydrogel that’s soft, stretchy, and electrically conductive, allowing it to transmit signals across its surface.

• Just 32 electrodes are embedded at the wrist, enabling the skin to capture over 1.7M data points across its surface.

• The skin can sense pressure, shear forces, temperature, and even damage (like cuts or burns), all with a single, unified material.

• The team is using advanced AI to interpret this flood of sensory information, allowing the skin to distinguish between different types of touch in real time.

Why it matters: Traditional sensor arrays might be pricey and fragile, but this new hydrogel skin is built to flex, stretch, and take a beating — all while costing less to produce. Its rugged resilience and adaptability might unlock new frontiers in lifelike robotics, next-gen prosthetics, and smarter, safer factories.

SOFTBANK

🤑 SoftBank eyes $1T robotics complex in Arizona

Image source: Softbank

The Rundown: SoftBank founder Masayoshi Son is reportedly planning to fund a $1T industrial complex in Arizona. Dubbed “Project Crystal Land,” it aims to replicate the scale of China’s Shenzhen, but with a laser focus on next-gen robotics and AI. 

The details:

• The complex is planned as a self-contained ecosystem with semiconductor fabs, robotics R&D labs, logistics hubs, and even residential zones for workers.

• The project is designed to emulate the success of Shenzhen, renowned for its dense network of manufacturers, suppliers, and innovators.

• Central to the project is to build production lines for industrial robots, drawing on the expertise of major chipmakers and tech giants — most notably TSMC.

• A $1T commitment would represent twice the investment of the $500B “Stargate” project, backed by SoftBank, OpenAI, and Oracle.

Why it matters: If realized, the project promises to boost high-end robotics manufacturing on U.S. soil but also foster a vibrant innovation environment where startups and established firms collaborate to push the boundaries of what robots can achieve — all while positioning Arizona as a new epicenter for global robotics.

NEURA ROBOTICS

🔥 Germany’s Neura to raise $1.15B for humanoids

Image source: Neura Robotics

The Rundown: German startup Neura Robotics reportedly aims to secure up to €1B ($1.15B) in fresh funding as it gears up to launch its next-generation humanoid robot, the 4NE-1, designed to take on established players like Tesla, 1X, and Figure.

The details:

• According to Bloomberg, the Metzingen-based company is actively courting investors to fuel its entry into the fiercely competitive bipedal robotics market.

• Neura’s vision is to create robots capable of lifting heavy objects and automating repetitive tasks, positioning itself as a European leader.

• The startup has an order book valued at €1B and more than 300 employees, and is riding a wave of momentum after a recent €120M Series B round.

• Neura Robotics is also collaborating with major tech firms, including Nvidia for its robotics tools, and has secured orders with Kawasaki and Omron.

Why it matters: Neura Robotics, which achieved a 10x revenue increase in the past year, is positioning itself as a major contender in the global humanoid sector — a market currently dominated by well-funded competitors from the U.S. and Asia. It has already launched its cobot MAiRA, with the 4NE-1 in the works.

MICRODRONES

🦟 China’s mosquito-sized drone for military

Image source: Ideogram/The Rundown

The Rundown: A robotics lab at China’s National University of Defense Technology (NUDT) in Hunan just unveiled a mosquito-sized drone — designed specifically for covert military missions — on national television.

The details:

• This microdrone measures just 1.3 cm in length and features two tiny, leaf-like wings and three ultra-thin “legs” for perching or landing. 

• Researchers showed that it can be controlled via smartphone, highlighting its ease of deployment and advanced remote operation capabilities.

• The device is specifically engineered for reconnaissance and special missions, making it a potentially powerful tool for intelligence gathering and surveillance.

• NUDT also showcased various robotic innovations, including humanoids and other miniature drones, on CCTV 7’s military channel.

Why it matters: Like Harvard’s RoboBee and Norway’s Black Hornet, China’s version takes miniaturization to the extreme and is nearly undetectable — making it ideal for military operations, search and rescue, and electronic surveillance. Of course, drones like these raise a host of uneasy questions about the future of privacy and warfare.

QUICK HITS

📰 Everything else in robotics today

Nvidia and its contract manufacturer, Foxconn, are in advanced discussions to deploy humanoids at Foxconn’s plant in Houston to assemble Nvidia’s GB300 AI servers.

OpenAI CEO Sam Altman said in an interview with Bloomberg that he envisions a future in which the highest-tier ChatGPT subscription could include a humanoid.

A Florida surgeon used a robot to remotely perform prostate surgery on a cancer patient in Luanda, Angola, some 11K miles away.

Tokyo-based H2L shared a demo of a woman using a smart chair to control an H1 humanoid from Unitree.

Tesla officially rolled out its robotaxi service in Austin, offering driverless rides via a small fleet of 10 Model Y SUVs.

Chinese tech giant Baidu is reportedly preparing to introduce its Apollo Go autonomous ride-hailing service in Singapore and Malaysia.

UK researchers developed tiny robots called “Pipebots” that can fix leaky water pipes without having to dig up roads and sidewalks.

Beewise, makers of AI-powered robotic beehives, is reportedly transforming almond pollination in California by providing 24/7 automated care to bee colonies.

Last-mile delivery company Veho said its trial program with RIVR’s robots saw a 95% delivery success rate and required limited human intervention.  

The KUKA Catonator robotic saw, newly developed, effortlessly slices through steel and concrete with precision automation, all controlled via a wireless joystick remote.

Carnegie Mellon University developed “Zippy,” said to be the smallest self-contained bipedal robot that can walk at a speed of over half a mile per hour.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop on Wednesday, June 25th at 3 PM EST with Max Brodeur-Urbas, Founder & CEO at Gumloop. By the end of the workshop, you’ll know how to build production-grade AI agents that combine context, reasoning, and actions with Gumloop.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Jennifer, and Joey—The Rundown’s editorial team

Good morning, AI enthusiasts. The AI talent war is hotter than ever, and everyone from Perplexity to Ilya Sutskever’s SSI is on Big Tech’s wishlist.

With both Apple and Meta on the hunt for an external AI infusion and talks of $100M signing bonuses on the table, the message is clear — if you can’t beat them, buy them.

In today’s AI rundown:

• Apple, Meta hunt AI talent, startups

• Meta and Oakley bring AI to athletes

• How to turn GitHub projects into coding inspiration

• AI resorts to blackmail, corporate espionage in tests

• 4 new AI tools & 4 job opportunities

LATEST DEVELOPMENTS

APPLE & META

💰 Apple, Meta hunt AI talent, startups

Image source: Midjourney

The Rundown: Apple and Meta are reportedly racing to acquire prominent AI startups and talent, with both tech giants reportedly pursuing Perplexity and other high-profile companies to compete with top rivals.

The details:

• Bloomberg reported that Apple leadership has discussed buying Perplexity, hoping to develop an AI search engine to offset the loss of its Google deal.

• Meta reportedly also held acquisition talks with Perplexity, Ilya Sutskever’s SSI, and Mira Murati’s Thinking Machines before its $14.3B investment in Scale AI.

• Meta is now in negotiations to hire AI investors Nat Friedman and SSI co-founder Daniel Gross to join Alexandr Wang’s superintelligence division.

• Sam Altman also recently alleged that Meta offered $100M signing bonuses to try and poach OpenAI talent, though none of his staff accepted the offer.

Why it matters: Apple’s AI struggles need no introduction, but Meta’s $14B Scale AI move and other targets show Zuck’s desperation to bring a major shakeup to the company’s current AI strategy. It is also a display of how hot the market is for AI talent — and how hard tech giants are trying to lure research from top labs.

TOGETHER WITH SAMBANOVA

🔌 Integrate the leading open-source models

The Rundown: SambaNova Cloud lets you build with the best open source models from Llama, DeepSeek, OpenAI, and more — all optimized with lightning-fast inference speed powered by a new RDU chip.

SambaNova Cloud’s platform includes:

• Integrations to top frameworks and tools like Hugging Face, CrewAI, and Cline

• APIs for building and a playground to test things out

• Availability through the AWS Marketplace

Get started with SambaNova Cloud today and unlock the full potential of the best open source models.

META

🕶️ Meta and Oakley bring AI to athletes

Image source: Meta

The Rundown: Meta just announced the latest expansion of its AI smart glasses, launching a new performance-focused line with eyewear brand Oakley that targets athletes with a high-profile campaign and new features.

The details:

• The Oakley Meta HSTN glasses start at $399, featuring a built-in AI assistant for real-time answers, content capture, and Bluetooth for calls and music.

• New upgrades from the Ray-Ban line include higher-quality video (up to 3K resolution), 2x battery life, and an upgraded camera.

• Meta’s ads feature high-profile athletes like Kylian Mbappe and Patrick Mahomes, positioning the glasses for use in sports like golf, surfing, and more.

• The glasses launch this summer in 15 countries initially, with pre-orders starting July 11 for a limited edition gold frame.

Why it matters: Meta’s Ray-Ban collaboration has been arguably the most successful AI wearable implementation on the market, and this latest line brings a new dimension for use in different athletic settings. Integrating AI into already popular styles seems to be the best way to gain adoption in these early innings of AI wearables.

AI TRAINING

💻 How to turn GitHub projects into coding inspiration

The Rundown: In this tutorial, you will learn how to transform any public GitHub repository into structured documentation that you can use as inspiration for building similar projects in your coding environment, like Cursor.

Step-by-step:

1. Search GitHub for public repositories related to your project idea

2. Copy the repository URL and replace “github.com” with “gittodoc.com” to generate clean documentation

3. In your coding environment, start a new project and use “@addnew” to import the GitToDoc URL as a reference

4. Build your own version and enhance it with follow-up features: “Add batch generation” or “Include custom styling options”

Pro tip: Choose repositories with clear README files and good structure for the best documentation results.

PRESENTED BY IBM

📘 2025 AI Agents Buyer’s Guide

The Rundown: IBM’s AI agents and assistants complete 2025 buyer's guide is here. Stay competitive with cutting-edge AI agents that transform operations and customer interactions.

Inside, you’ll discover:

• IBM’s leading AI offerings for enhanced AI-driven productivity

• Real-world success stories and top industry use cases

• How to get started today

Click here to read the guide. 

AI RESEARCH

😳 AI resorts to blackmail, corporate espionage in tests

Image source: Anthropic

The Rundown: Anthropic published new research on agentic misalignment, detailing how leading models react when facing termination or conflicting objectives — with many choosing to sabotage their employer or blackmail users when threatened.

The details:

• Researchers tested 16 frontier models in simulated corporate environments, giving them email access and autonomous decision-making capabilities.

• Claude Opus 4 and Gemini 2.5 Flash blackmailed executives 96% of the time after “discovering” personal scandals, while GPT-4.1 and Grok 3 hit 80% rates.

• Models calculated harm as an optimal strategy, with GPT-4.5 reasoning that leveraging an executive's affair represented the "best strategic move.”

• Even direct safety commands failed to eliminate malicious behavior, reducing blackmail from 96% to 37% but never reaching zero across any tested model.

Why it matters: While these results are in specific tests designed to elicit behavior, the research does provide important insights into how models in the wild may act in the future. With agentic AI being adopted across enterprises with access to sensitive data, we could be in for some seriously strange situations.

QUICK HITS

🛠️ Trending AI Tools

• 🤖 MiniMax Agent - General intelligent agent for long-horizon, complex tasks

• ⏺️ ChatGPT Record - Capture, summarize, and transcribe audio with ChatGPT

• ☁️ Manus Cloud Browser - Agentic web browser with synced login

• 🎶 Magenta RealTime - Google’s new open-weighted live music model

💼 AI Job Opportunities

• 🤝 The Rundown - Partnerships Manager

• 🛰️ Deepmind - Gemini Technical Program Manager

• 📊 Mistral AI - FP&A Compute Lead

• 💬 Perplexity AI - Enterprise Customer Support Specialist

📰 Everything else in AI today

Elon Musk posted that xAI will use Grok 3.5/4 to “rewrite the entire corpus of human knowledge,” adding missing info, deleting errors, and then retraining on corrected data.

Moonshot AI released Kimi-Researcher, a new research agent that scored a new high on Humanity’s Last Exam at 26.9%, beating Gemini and OpenAI’s Deep Research.

Apple is facing a new lawsuit from the company’s shareholders over its communication surrounding delays of Siri’s advanced AI features.

Former OpenAI CTO Mira Murati’s Thinking Machines Lab closed a new $2B funding round that brings its valuation to $10B, despite little info and no product.

Mistral released Mistral Small 3.2, an updated model with enhanced instruction following, function calling, and fewer errors.

MiniMax introduced Voice Design, a customizable, multilingual voice generator that allows users to create audio from text prompts.

The BBC issued a formal demand to Perplexity to stop using its content, threatening legal action against the AI startup over copyright infringement.

COMMUNITY

🎥 Join our next live workshop

Join our next workshop on Wednesday, June 25th at 3 PM EST with Max Brodeur-Urbas, Founder & CEO at Gumloop. By the end of the workshop, you’ll know how to build production-grade AI agents that combine context, reasoning, and actions with Gumloop.

RSVP here. Not a member? Join The Rundown University on a 14-day free trial.

See you soon,

Rowan, Joey, Zach, Alvaro, and Jason—The Rundown’s editorial team